摘要
为更合理和准确的应用经验模态分解EMD方法,我们通过数值实验全面的研究了不同采样率对EMD的影响。结果表明采样率从两个方面影响经验模态分解,其一是由于离散化带来的非对准,导致的信号非对称从而给本征模态函数的提取带来的能量泄漏,其二为过高的归一化频率带来信号表达的粗糙化而引起的分解结果的正交性降低。实验结果也表明:较为理想的EMD应用环境应该为采样频率大于4倍的Nyqvist频率,而应用EMD的最低要求应该是:采样频率大于2倍的Nyqvist频率。该结果将对EMD方法的工程应用有着重要的指导意义!
The aim of this paper is to investigate the effect of sampling on the empirical mode decomposition （EMD）. To this end,numerical experiments have been employed to analyze the error sources of EMD due to the sampling. The results show that there are two kinds of errors are introduced by the sampling procedure. The first one is attributed to the sampling instant of the discretization not aligning with the local maximum of the waveform and then introduces energy leakage, and the other is attributed to poor amplitude resolution of the inspected signal due to lower sampling rate then introduces decrease of orthogonality. The results also led to the conclusion that a sampling rate which was more than 4 times the Nyqvist rate will conduce to preferable performance of the EMD, and the sampling rate of more than 2 times the Nyqvist rate is needed for the normal application of EMD. This conclusion will play very important role in the EMD application.
出处
《信号处理》
CSCD
北大核心
2007年第4期637-640,共4页
Journal of Signal Processing
基金
广东自然科学基金(No:05006593)
